How many flux towers are enough? Energy balance closure and large eddy simulation as diagnostic tools for secondary circulations

The majority of eddy covariance towers fail to close the surface energy balance, with turbulent and storage fluxes of heat and water smaller than the available energy. But, this closure appears to get worse the greater the spatial heterogeneity in land surface properties. Previous studies have suggested this effect relates to the presence of energy-transporting mesoscale secondary circulations that can not be sampled by a single flux tower. It is possible that multiple flux towers within a small domain, by conducting “spatial” eddy covariance and sampling across heterogeneity, might be able to resolve this circulation and improve both energy balance and model-data comparison. To test this hypothesis, we preformed initial large eddy simulation experiments in a flat but spatially mixed cover domain to test whether increasing heterogeneity in surface energy fluxes induces organized circulations and to identify the minimum number of flux towers needed to evaluate that circulation. These findings will be used to prepare experimental design and diagnosis of energy balance closure for the proposed Chequamegon Heterogeneous Ecosystem Energy-balance Study Enabled by a High-density Extensive Array of Detectors (CHEESEHEAD) intensive field-campaign. The large flux tower array campaign would be based in Northern Wisconsin and aims to address long-standing puzzles regarding the role of atmospheric boundary-layer responses to scales of spatial heterogeneity in surface-atmosphere heat and water exchanges.

The LES we used was the System for Atmospheric Modelling (SAM). The simulation was setup in a 10x10 km domain with 40x40 meter horizontal resolution with 128 vertical levels and doubly periodic boundary conditions, and run during a period where tower-derived scaled fluxes were made available using the Environmental Response Function technique. A spatial surface heterogeneity was created by sinusoidally varying the Bowen ratio throughout the domain scaled to match the pattern in observations.  We deployed increasing numbers of “virtual” flux towers in the landscape, computing both temporal-based and spatial-based energy fluxes. We hypothesize energy balance heterogeneity drives development of circulations, requiring a significant number of towers to sufficient resolve and sample the mean domain energy flux.